Oklahoma DOT Perpetual Pavement Test Sections at the NCAT Test Track

Compression Tension

Perpetual Pavement Structure and materials are designed to provide a long service life (50 years +)

Early Texas Perpetual Pavements Waco Cotulla McAllen 1.5” PFC 2.0” SMA 3.0” 19 mm SP 10.0” 25 mm SP 4.0” 19 mm 2% air voids 6.0” crushed stone 3.0” SMA 3.0” 19 mm SP 8.0” 25 mm SP 2.0” 12.5 mm 2% air voids 1.5” PFC 2.0” SMA 10.0”, 19 mm or 25 mm SP 3.0” 12.5 mm 2% air voids 8.0” lime-treated salvaged aggregate

Waco Cotulla McAllen 16” Total HMA 20.5” Total HMA 16.5” Total HMA 1.5” PFC 3.0” SMA 1.5” PFC 2.0” SMA 2.0” SMA 3.0” 19 mm SP 3.0” 19 mm SP 10.0”, 19 mm or 25 mm SP 8.0” 25 mm SP 10.0” 25 mm SP 3.0” 12.5 mm 2% air voids 2.0” 12.5 mm 2% air voids 4.0” 19 mm 2% air voids 8.0” lime-treated salvaged aggregate 16” Total HMA 6.0” crushed stone 20.5” Total HMA 16.5” Total HMA

How thick is too thick? Cost of 5 Miles of Pavement Assume 80’ width, $50 per ton Save 1” in over-design: $650,000 Save 2” in over-design: $1,300,000 Save 4” in over-design: $2,600,000

How thin is too thin? If the perpetual pavement structure is designed too thin, the risk of bottom-up cracking increases dramatically, defeating the purpose of the design and resulting in an expensive rebuild.

Rut Resistant Material To design the appropriate pavement thickness, stresses and strains To design against potential bottom-up cracking, certain strain thresholds cannot be exceeded. } 1.5-3 in. PFC, SMA, etc. 3” to 6” High Shear Zone High Modulus Rut Resistant Material (Varies As Needed) Max Flexural Strain Pavement Foundation

Limiting Strain Theory Based on laboratory beam fatigue data “Small” strains will never induce cracking Limiting strain at 75 to 125 microstrains No bottom-up cracking below this level

These strains can be calculated based on total pavement thickness and material properties. However, the calculations would be based on assumptions made from previously collected data.

There are large deviations in the types of environment around the country, the types of materials used and the types of pavement specified. These deviations decrease the reliability of the assumptions made to calculate strains. Estimated Strain If the strains could be measured directly, using Oklahoma materials and mix designs, the data would be much more reliable.

National Center for Asphalt Technology (NCAT) Test Track

ODOT’S PERPETUAL PAVEMENT STRUCTURAL SECTIONS AT NCAT TEST TRACK PLAN VIEW SECTION N8 – 150’ SECTION N9 – 150’ 25’ TRANSITION 25’ TRANSITION 50’ TRANSITION

ODOT’S PERPETUAL PAVEMENT STRUCTURAL SECTIONS AT NCAT TEST TRACK PLAN VIEW SECTION N8 – 150’ SECTION N9 – 150’ 25’ TRANSITION 25’ TRANSITION 50’ TRANSITION PROFILE VIEW 2” SMA w/PG 76-28 3” SuperPave 19.0mm w/PG 76-28 3” SuperPave 19.0mm w/PG 64-22 2” RBL w/PG 64-22 3” SuperPave 19.0mm w/PG 64-22 3” RBL w/PG 64-22 *RBL = RICH BOTTOM LAYER

Structural Study Test Sections N8 and N9 are part of an eleven-section structural experiment starting in 2006. FHWA has partnered with OK on section N9 to have additional instrumentation installed in section N9 to more precisely measure strain and temperature at various depths. There is an error on this graphic - please note that the Missouri Type 5 base is only 4 inches thick.

Rut Resistant Material The maximum flexural strain is directly measured using a series of strain gauges installed at the bottom of the asphalt section. 1.5-3 in. PFC, SMA, etc. High Modulus Rut Resistant Material (Varies As Needed) Max Flexural Strain Pavement Foundation

If we have determined the strain for a 10” and 14” thick pavement, the thickness at the critical strain level can be interpolated / extrapolated ? MICROSTRAIN 10” 14” 16”

Instrumentation – N8 12 Asphalt Strain gauges (6 longitudinal, 6 transverse) 2 pressure plates (1 at top of base, 1 at top of subgrade) Laser to measure wheel wander Temperature array to measure temps at top, middle and bottom of HMA

Accelerated Loading Compress 10-15 years into 2 years Heavy triples optimize safety and efficiency Each axle loaded to federal legal bridge limit Heavier loads induce unrealistic damage No bridge spans to overload on NCAT track Average gross vehicle weight 155k pounds

Long Term Response Measurements Weekly - three passes of 5 trucks Steer axle (12 kip), tandem (40 kip), 5 single axles (20 kip/axle) 45 MPH Measure strain Measure pressure Measure temps Characterize long-term trends in strain response Graph shows transverse strain in section N9 for one truck pass. The steer, tandem and 5 single axles are clear to see.

Graph shows transverse strain in section N9 for one truck pass Graph shows transverse strain in section N9 for one truck pass. The steer, tandem and 5 single axles are clear to see.

To design the appropriate pavement thickness, stresses and strains The vertical compressive strain is directly measured using pressure plates installed at the top of the subgrade. 1.5-3 in. PFC, SMA, etc. High Modulus Rut Resistant Material (Varies As Needed) Max Flexural Strain Vertical Compressive Strain

Moisture sensors were also placed in the subgrade to continuously monitor soil conditions

Temperature probes were placed in the pavement to continuously monitor temperature at various depths

A datalogger is used to collect information from the different pavement sensors. In addition to the normal “slow speed” mode that continuously gathers data, a “high speed” mode can be used to collect 40,000 data points per second.

Additional Testing - Surface Map Cracking

FWD Testing Rut Testing Skid Testing Inertial Profiler –. Rutting FWD Testing Rut Testing Skid Testing Inertial Profiler – Rutting Smoothness Surface Texture

N8 and N9 – Strain vs. Date This plot shows the 95th percentile measured longitudinal strain from single axles in sections N8 and N9. In a given day’s worth of data collection, we have 15 truck passes * 5 single axles per truck = 75 measurements. The data on this plot represent the 95th percentile of the 75 measurements. As expected, N9 is significantly lower than N8. Ratios were computed by dividing N8 strain by N9 strain. The ratio varies somewhat, but on average, the strain in section N8 is about 2.6 times greater than N9. The seasonal trend is also evident in both sections where the strains rise significantly in the warmer summer months.

Effect of Depth (N9) This plot shows the effect of depth on strain within section N9. These are measured longitudinal strains under single axles. As expected, the bottom of the HMA sees the greatest strain since it’s at the extreme point going through the most bending. As you move up in the structure, the strains are reduced since they are approaching the neutral axis where the strain is zero. An important consideration here is that each strain plot represents different mixtures. Theoretically, they may have different strain thresholds. So, even though the strains are reduced, they may not be as fatigue resistant compared to the rich bottom. Note: There are gauges in Lift 4, however after reviewing the data, they were not included. They were very erratic which was not unexpected since they are so shallow.

Temperature Normalized Response Section N9 is the only section in the structural experiment that could be classified as perpetual based on these average temperature normalized responses; however, the analysis is ongoing.

Normal Operations 55% of Strains below threshold The measured mid-depth HMA temperature was used with the speed and temperature equation to determine strains between the start of the experiment (11/10/06 and 1/14/2008). As shown in the graph, and shown previously for both N8 and N9, the seasonal trend is clearly evident. Also, strains in N9 can be quite high. For example, the highest measured temperature was on August 9, 2007 (118F) between 6 and 7 PM. During this hour, strains were on the order of 580 microstrain. Again referring to a fatigue threshold of 100 microstrain, the graph shows that approximately 55% of the strains were below the threshold with 45% exceeding the threshold. Until more performance data are recorded, it is difficult to ascertain whether 100 microstrain is appropriate to use as the fatigue threshold. What can be said, however, is that a significant number of strain readings exceed 100 microstrain. It is important to note the difference in how this graph was generated compared to the N8, N9 graph shown previosly. The previous graph was generated from measured data only. This graph was generated by developing a strain vs. temperature equation from which we could reasonably approximate strains during any hour of testing.

Supplemental FHWA Experiment N9 Instrumentation (Multi-Depth Array) 12 strain gauges at bottom of HMA 2 pressure plates (top of base, top of subgrade) Additional three groups of four gauges/group measuring strain on top of successive paving lifts Laser to measure wheel wander

N9 Cross Section and Supplemental Instrumentation Additional instrumentation funded by FHWA as part of strain vs. speed and temperature study Instrumentation same as N8, plus gauges installed on successive lifts going up into the structure. More detailed temperature readings at top, middle and bottom of each paving lift

Strain vs. Speed and Temperature (ODOT and FHWA) Primary Objectives Study a perpetual pavement under a wider set of conditions Primarily look at speed and temperature effects

Data Collected 4 Speeds Tested 4 Dates Measured 15-45 mph (10 mph increments) 4 Dates April – May 2007 Measured Temperature Tensile Strain

Investigations Studied relationships Strain Threshold Load Durations Speed-Strain Temperature-strain Combined Strain Threshold Load Durations Relate to Frequency

Relationships Speed-Strain Temperature-Strain Strain at bottom of HMA Decrease with speed Temperature-Strain Strain at bottom of HMA Increase with temperature Trends were examined that looked first at speed effects and secondly the temperature effects in N9. The first equation shows that strain is related to the natural log of speed and tends to decrease as speed is increased. The second equation shows that strain tends to increase with temperature. A third equation was developed that incorporates both variables. The R2 on the data set was very high (0.955-0.986) R2: 0.955-0.986

Strain Threshold Tensile strain threshold: 100με These design curves show the effect of temperature on strain at a variety of reference speeds. For example, if the design speed were 45 mph, the critical temperature would be approximately 72F. Above this temperature, strains would exceed the commonly held fatigue threshold of 100 microstrain. At 65 mph, the critical temperature would be approximately 80F. Curves such as these can help to extrapolate the findings from the test track to those encountered on open access facilities (i.e., in Oklahoma).

THANK YOU! QUESTIONS?